Housekeeping/Backup

[seravat]

Hi guys,
This is not an issue but more some questions.

So I have installed the operator on ocp 3.11 to generate some useful reports. The cluster is running on azure with: 3 worker nodes, 3 infra, 3 dedicated for prometheus and 3 masters and 3 for ocs. The scaling for the number of nodes shouldn't be huge in the future.

In this first test installation I have used pv's as storage for both the hive warehouse and the metastore, which I believe it is fine for this setup.

Does the operator do any housekeeping - e.g. deleting old records and if yes, what is the configuration for that and can it do backups??

Also, do you feel this setup is "recommended"?

Cheers,
[email protected]

[chancez]

We don't do any housekeeping, we typically expect to keep the data forever since we store it in object storage and for many uses you want to be able to query historic data. It's something we may consider adding in the future.

As far as backups, this arguably something the platform could provide. The only data you would need to back up is the Hive metastore PVC, and I believe there are many existing ways to do backups, and as far as Metering is concerned, it just wants the PVC to have the correct stuff. If you're backing up PVCs with something, and metering has issues, you could re-create the PV/PVC for the Hive from a backup without Metering being involved for the most part.